I have a system where I have a NIOS running at 100MHz and an SDRAM running at 100Mhz.
When I do some profiling to see how fast my memory throughput to the CPU is, I am getting about 700,000 accesses per second. Not even 1% of the 100Mhz system clock.
I am using the altera SDRAM controller and I was really hoping to leverage this IP so that I didn't have to write my own SDRAM controller. Is it really that slow?
Anyone have any idea of what the throughput would be if a custom master was connected to this IP and trying to push data into the memory as fast as possible?
I am wanting to hook up this SDRAM controller to my own master that will need to be able to push data into the memory at speeds up to 6MHz. I had assumed the SDRAM running at 100MHz would get the job done, but now I'm a bit worried. Maybe its just because the CPU is in charge? Also, it seems the altera SDRAM controller does not support bursting either.
